As the pixel size of complementary metal oxide semiconductor (CMOS) image sensors becomes smaller for higher pixel density and lower cost, the active area of the photodetector also becomes smaller. For pinned photodetectors that are commonly used in CMOS image sensors, the smaller photodetector area leads to reduced full-well capacity, meaning that the maximum number of charges that can be held in the photodetector is reduced. The reduced full-well-capacity in turn results in a pixel with lower dynamic range and lower signal-to-noise ratio. Therefore, methods to increase the full-well-capacity of the pinned photodetector are highly desired.
In the p-n-p pinned photodetector most commonly used for CMOS image sensors, the most straightforward way to increase the pixel's full well capacity is to increase the doping level (i.e., the concentration of dopants) in the n-type layer, for example by increasing the implantation dosage. For small pixel sizes, however, the increased n-type doping can lead to significant increase in dark current and in defective pixels commonly referred to as white pixels. One reason for this is because of the increased electrical field along shallow trench isolation (STI) sidewalls due to the high n-type doping and the shrinking distance between n-type implant and STI edge.